國立高雄大學圖資館 |

Language: English

Back

使用代理曲面的多階無微分最佳化方法 = Multiple Stage A...

國立高雄大學應用數學系碩士班

使用代理曲面的多階無微分最佳化方法 = Multiple Stage Approaches in Surrogate Assisted Derivative Free Optimization

Record Type:	Language materials, printed : monographic
Paralel Title:	Multiple Stage Approaches in Surrogate Assisted Derivative Free Optimization
Author:	曾詩婷,
Secondary Intellectual Responsibility:	國立高雄大學
Place of Publication:	[高雄市]
Published:	撰者;
Year of Publication:	2009[民98]
Description:	101面圖、表 : 30公分;
Subject:	代理曲面
Subject:	derivative free
Online resource:	http://handle.ncl.edu.tw/11296/ndltd/66467234798594406173
Notes:	參考書目：面
Notes:	指導教授：王偉仲
Summary:	在電腦模擬實驗過程中，最佳化問題的反應函數通常假定是未知的。因此，找到最佳解是很困難的。數種區域最佳化方法已經被提議而沒有使用反應函數的微分資訊。然而，這些區域最佳化方法只能找到單一極值，因此如何找到多重極值是一個很重要的議題。在這篇論文裡，我們的目標是根據更新過代理曲面找到有效的起始點。首先，在每一個階段先構造一個代理曲面來估計真實且未知的反應曲面，然後根據代理曲面，用一個機制篩選有效的起始點來結合區域最佳化方法。幾個數值實驗顯示我們的方法是有效的。 The paper functions for optimization problems in the computer experiments are usually to be unknown. Hence it is difficult to find optimal points. Several local optimization methods have been proposed without using derivative information of the resp onse function. However, these local methods can only be used to identify one local extreme; hence how to find multiple local extremes is an important issue. In this thesis, our goal is to get the efficient starting points from the improved surrogate surface for local optimization method. Here we construct the surrogate surface in each stage to a pproximate the true but unknown response surface, and then base on this surrogate surface, the starting points are selected by a screening procedure for local optimizers. Numerical experimental results are showed that our methodperforms well.

使用代理曲面的多階無微分最佳化方法 = Multiple Stage Approaches in Surrogate Assisted Derivative Free Optimization
曾, 詩婷

使用代理曲面的多階無微分最佳化方法 = Multiple Stage Approaches in Surrogate Assisted Derivative Free Optimization / 曾詩婷撰 - [高雄市] : 撰者, 2009[民98]. - 101面 ; 圖、表 ; 30公分.
參考書目：面指導教授：王偉仲.
代理曲面derivative free

使用代理曲面的多階無微分最佳化方法 = Multiple Stage Approaches in Surrogate Assisted Derivative Free Optimization
LDR:02610nam a2200277 4500 001 220431
005 20170214093134.0
009 220431
010 0 $b 平裝
010 0 $b 精裝
100 $a 20170214y2009 k y0chiy09 ea
101 1 $a chi $d chi $d eng
102 $a tw
105 $a ak am 000yy
200 1 $a 使用代理曲面的多階無微分最佳化方法 $d Multiple Stage Approaches in Surrogate Assisted Derivative Free Optimization $z eng $f 曾詩婷撰
210 $a [高雄市] $c 撰者 $d 2009[民98]
215 0 $a 101面 $c 圖、表 $d 30公分
300 $a 參考書目：面
300 $a 指導教授：王偉仲
328 $a 碩士論文--國立高雄大學應用數學系碩士班
330 $a 在電腦模擬實驗過程中，最佳化問題的反應函數通常假定是未知的。因此，找到最佳解是很困難的。數種區域最佳化方法已經被提議而沒有使用反應函數的微分資訊。然而，這些區域最佳化方法只能找到單一極值，因此如何找到多重極值是一個很重要的議題。在這篇論文裡，我們的目標是根據更新過代理曲面找到有效的起始點。首先，在每一個階段先構造一個代理曲面來估計真實且未知的反應曲面，然後根據代理曲面，用一個機制篩選有效的起始點來結合區域最佳化方法。幾個數值實驗顯示我們的方法是有效的。 The paper functions for optimization problems in the computer experiments are usually to be unknown. Hence it is difficult to find optimal points. Several local optimization methods have been proposed without using derivative information of the resp onse function. However, these local methods can only be used to identify one local extreme; hence how to find multiple local extremes is an important issue. In this thesis, our goal is to get the efficient starting points from the improved surrogate surface for local optimization method. Here we construct the surrogate surface in each stage to a pproximate the true but unknown response surface, and then base on this surrogate surface, the starting points are selected by a screening procedure for local optimizers. Numerical experimental results are showed that our methodperforms well.
510 1 $a Multiple Stage Approaches in Surrogate Assisted Derivative Free Optimization $z eng
610 0 $a 代理曲面 $a 多重 $a 最佳化 $a 無微分
610 1 $a derivative free $a multiple $a optimization $a surrogate surface
681 $a 008M/0019 $b 462101 8004 $v 2007年版
700 1 $a 曾 $b 詩婷 $4 撰 $3 354093
712 0 2 $a 國立高雄大學 $b 應用數學系碩士班 $3 166142
801 0 $a tw $b 國立高雄大學 $c 20091020 $g CCR
856 7 $2 http $u http://handle.ncl.edu.tw/11296/ndltd/66467234798594406173